Research And Implementation Of Network Caching System In Software-Defined Networking Environment Based On Openflow

With the rapid development of Internet services, Internet traffic is increasing significantly. When users read online news, enjoy online video or watch IPTV, the popular content are always requested repeatedly, which takes up a large amount of network bandwidth resource and cause the drop of the quality of network service.Exiting solutions mainly adopt application-layer technologies to increase the quality of network service, such as CDN (Content Distribution Network) and transparent proxy technology, which alleviate the traffic pressure in the Internet to some extent. However, there are several deficiencies by deploying the service overlay the application-layer of the network. The content providers can not realize the running status of the network and modify the caching strategy dynamically with transparent cache or reverse proxy technology. Besides, the cache services are unable to provide customized services for end users without the information of global network and dynamical control of data routing.To overcome the existing problem of network cache solutions, this article proposed a network caching system based on OpenFlow. By taking the advantage of the centralized control of data transmission and the support of network visualization in SDN environment, we use the Floodlight controller to handle the content request in network and achieve the flexible management of content caching according to the network status, virtual networks priority and the user level, which makes the system able to provide more efficient cache strategies and customized services. Furthermore, by writing the northbound interface in Floodlight controller, a web interface is provided to achieve visual operation for the management of the proposed system.At last, we build an SDN experimental environment using Open vSwitch. And we have evaluated the cache services for users in different virtual networks and users with different levels in a same virtual network respectively. The results show that our new network cache system have achieved the design targets and successfully provided customized services for different users.